The term real-time is meaningless unless you qualify it with a deadline and an acceptable miss rate. When you know that you can determine whether or not Linux is a suitable candidate for the operating system and, if so, begin to tune your system to meet the requirements. Tuning Linux and your application to handle real-time events means making it more deterministic so that it can process data reliably inside deadlines. Determinism usually comes at the price of total throughput so a real-time system is not going to be able to process as much data as a non-real-time system.

It is not possible to provide mathematical proof that a complex operating system like Linux will always meet a given deadline, so the only approach is through extensive testing using tools such as cyclictest and Ftrace, and, more importantly, using your own benchmarks for your own application.

To improve determinism, you need to consider both the application and the kernel. When writing real-time applications, you should follow the guidelines given in this chapter about scheduling, locking, and memory.

The kernel has a large impact on the determinism of your system. Thankfully, there has been a lot of work on this over the years. Enabling kernel preemption is a good first step. If you still find that it is missing deadlines more often than you would like, then you might want to consider the PREEMPT_RT kernel patches. They can certainly produce low latencies but the fact that they are not in mainline yet means that you may have problems integrating them with the vendor kernel for your particular board. You may instead, or in addition, need to embark on the exercise of finding the cause of the latencies using Ftrace and similar tools.

That brings me to the end of this dissection of embedded Linux. Being an engineer of embedded systems requires a very wide range of skills, which range from a low level knowledge of hardware, how the system bootstrap works and how the kernel interacts with it, to being an excellent system engineer who is able to configure user applications and tune them to work in an efficient manner. All of this has to be done with hardware that is, almost always, only just capable of the task. There is a quotation that sums this up, An engineer can do for a dollar what anyone else can do for two. I hope that you will be able to achieve that with the information I have presented during the course of this book.